Incident Library

API · MCP 한국어

API and MCP kubegap Corrections and proposals

Incident Library

API · MCP 한국어

Component

Karpenter

Karpenter · karpenter

3 cases · 2024 to 2025

Root cause Capacity exhaustion2 Software defect1

Contributing factor Automation amplification2 Unverified assumption1 Unreviewed defaults1 Harmful manual action1 No capacity planning

Symptom Autoscaler misbehaviour1 GPU unavailable1 High latency1 Intermittent connection failures1

Organisation Adevinta1 Inflab1 이언철1

Cases, by severity

IL-0171 Karpenter node auto-repair replaced EKS GPU nodes about 10 minutes after they became Ready, in a loop, because the Bottlerocket NVIDIA AMI lacked libdcgm.so and set AcceleratedHardwareReady to FalseInflab · 2025
Root cause Software defect
IL-0007 The AWS VPC CNI could not assign IP addresses to pods as an EKS cluster ran out of IPv4 spaceAdevinta · 2024
Root cause Capacity exhaustion
IL-0182 Too few CoreDNS replicas on EKS were overloaded during traffic spikes, raising latency and cutting the network for task pods, and the over-provisioned, tightly spread CoreDNS pods added later blocked Karpenter consolidation

View on the graph

1

IP address exhaustion1

Node recreation loop1

이언철 · 2024

Root cause Capacity exhaustion